Hi, Apparently, you are in a "bona fide" twinning situations, and yet you can't detwin properly the data. Would it be possible that, while your crystal was rotating during data collection, different volumes of the two twin crystals were irradiated. Therefore, depending on the phi angle, the twinning fraction would have varied. Of course, in such a (hypothetical...) situation, merging everything would obscure the signal.
So one obvious questions: was your crystal fully bathed in the beam ? If not: would be interesting to try having a look at the unmerged data... Well, easier to suggest than to do... By the way, there may well exist papers on such problems. Philippe Dumas IBMC-CNRS, UPR9002 15, rue René Descartes 67084 Strasbourg cedex tel: +33 (0)3 88 41 70 02 [EMAIL PROTECTED] -----Message d'origine----- De : CCP4 bulletin board [mailto:[EMAIL PROTECTED] la part de Jordi Benach Envoyé : mardi 24 avril 2007 17:56 À : [email protected] Objet : [ccp4bb] MAD and twinning Hi, I just got a MAD dataset (~2.5A) of a crystal that so far seems to belong to P31 or P32 space group and that also scales pretty well in P3121 or P3221. According to the UCLA twinning server, the P31/P32 crystal is merohedrally twinned with a twinning fraction of 0.33 and following a twinning operation h,-h-k,-l. CNS gives a twinning fraction of 0.42. Although we have solved twinned crystals in the lab with similar twinning fractions/resolution in the past using SOLVE or BnP/SHELX, only realizing that the crystal was twinned during refinement when the R values wouldn't go down. In this P31/P32 case, however, I can't solve the twinned data with any program, or at least I can't get interpretable electron density maps. Any ideas, clues? Is there a way to get a program that will detwin raw intensity data (like *.sca files directly from scalepack)? Thanks, Jordi
